Memory Row Hammer Mitigation via Weak Row List
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As memory integration increases, the coupling effect between word lines leads to data damage in memory cells due to row hammering, where frequently activated word lines disturb adjacent cells, causing data loss before refresh operations can occur.
Innovation Solution
A memory system with a list storage circuit to identify vulnerable rows, a row hammer attack detection circuit to select and prioritize these rows for smart refresh operations, and a sampling circuit to increase the probability of selecting affected rows for refresh, thereby mitigating data loss from row hammering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the spacing between word lines is decreased to increase memory integration, then memory density is improved, but the coupling effect between neighboring word lines increases causing row hammering attacks
Solution Approach 1:
The patent identifies vulnerable rows in advance through a sampling circuit that monitors activation patterns and a weak row list storage circuit that pre-records rows susceptible to row hammering. During smart refresh operations, these pre-identified vulnerable rows are prioritized for refreshing before actual data corruption occurs, preventing the harmful effect rather than reacting to it after damage happens.
Solution Approach 2:
The patent implements a feedback mechanism where the sampling circuit continuously monitors word line activation patterns, feeds this information to the weak row list storage circuit, and updates the identification of vulnerable rows based on observed activation frequencies. This closed-loop feedback system dynamically adapts to usage patterns and adjusts refresh priorities accordingly.
2Reliability
If frequent refresh operations are performed on all rows to prevent data loss, then data integrity is improved, but memory operation efficiency deteriorates
Solution Approach 1:
The patent applies local quality by differentiating refresh operations between vulnerable rows and normal rows. Instead of uniform refreshing across all memory rows, the system applies enhanced refresh operations selectively to identified vulnerable rows while maintaining standard refresh operations for other rows. This localized approach concentrates protection resources where they are most needed while preserving overall system performance.
Solution Approach 2:
The patent implements partial action by performing smart refresh operations only on a subset of rows identified as vulnerable through the weak row list, rather than refreshing all rows. The sampling circuit captures a portion of active addresses and the selection circuit chooses specific rows for enhanced refreshing, applying excessive refresh action only where necessary to combat row hammering while minimizing impact on overall memory throughput.
3Measurement precision
If a sampling circuit is used to select rows for smart refresh, then the probability of selecting vulnerable rows is improved, but device complexity increases
Solution Approach 1:
The patent segments the row selection process into distinct functional modules: a sampling circuit that captures active addresses, a weak row list storage circuit that maintains vulnerable row information, and a selection circuit that integrates both data streams to identify target rows for smart refresh. This segmentation allows each component to perform its specific function efficiently while maintaining overall system manageability and clarity.
Data Source
AI summary
A memory includes: a memory core; a list storage circuit suitable for storing a weak row list of rows that are vulnerable to a row hammer attack in the memory core; and a row hammer attack detection circuit suitable for selecting rows that are row-hammer-attacked among rows in the memory core as hammered rows, and increasing a probability that the rows stored in the list storage circuit are selected as the hammered rows.